Your Black Friday campaign starts at midnight. The newsletter has gone out and the campaign page shows the correct discount. But the basket still displays the old price. Another customer has paid, yet their order remains ‘awaiting payment’. Where do you start looking?
This is an illustrative scenario, not an observed incident. It shows what needs organising before a promotion: the journey from advertising to an order. An offer becomes deliverable when pricing, stock, checkout and processing follow the same rules.
Turn the offer into rules your systems can enforce
‘Twenty per cent off selected products’ sounds clear until someone chooses a variant, adds an existing discount code or checks out just before the promotion ends. Decide which products qualify, which discounts can be combined and when the price is fixed. Specify start and end times with a time zone, and agree what happens to a basket filled before the deadline but paid for afterwards.
Enforce those rules in the backend. A crossed-out price in a theme is presentation; the final calculation must also be correct for a direct checkout request. Store the price used and the discounts applied with the order, so customer service can explain the total later.
For consumer promotions using former-price comparisons in the Netherlands, price history matters too. The Dutch regulator ACM states that the promotional price must generally be compared with the seller’s lowest price in the preceding thirty days. Exceptions apply, including for perishable goods and successive price reductions. The government’s Ondernemersplein guidance confirms the main rule. This paragraph describes the Dutch context; it is not a statement of the rules in every country.
The technical consequence is to retain price history for each relevant product variant and sales context. Fields holding only the current and ‘regular’ prices are insufficient for that purpose. Before launching, record the reference price for the offer and how it was determined. The discount badge can then be backed by traceable data.
Speed up the campaign without freezing the order
A campaign page with product images can often be served largely from a cache, avoiding rebuilding the same page for every visitor. But define when that stored version expires: both the beginning and the end of the discount need to become visible.
A basket poses a different problem because it contains customer-specific data. WooCommerce therefore recommends excluding Cart, Checkout and My Account from page caching; sessions and cookies also need attention in the chosen cache configuration. This is specific WooCommerce guidance, not a ban on every form of caching within an online store. See the official caching documentation.
In our scenario, check two paths: does a new visitor see the promotional price, and does an existing basket calculate that same promotion correctly? Repeat after the offer ends and with two separate user sessions. Test any hosting and CDN caches as well.
For more on that intermediary layer, our sponsor BlazeForce explains what Cloudflare adds and which cache checks remain necessary (in Dutch). Settings need to match the data a page displays.
The last item needs an explicit stock policy
Suppose two customers order the last item simultaneously. Periodic stock synchronisation alone does not resolve that collision. Choose which system determines sellable stock, and have it reserve or deduct inventory in a way that handles concurrent requests. The implementation depends on your commerce platform and inventory architecture.
Distinguish physical stock, reserved quantities and what remains available to sell. Agree when a reservation expires and what happens if a payment confirmation arrives afterwards. Without that decision, a technically successful payment can still require manual recovery.
Apply the same policy across campaign channels. Where possible, let advertisements or product feeds respond to availability, while allowing for delays between systems. With limited stock, a safety buffer may be more sensible than offering every physical unit online. Size it according to your sales channels and synchronisation delays; there is no universally correct percentage.
A payment does not end on the thank-you page
A customer pays and then closes the tab. If order processing starts only when the thank-you page opens, the order can stall. Stripe describes precisely this risk in its fulfillment documentation: the customer’s return to the site is not guaranteed. The server needs to process payment status independently.
With a webhook, the payment provider sends an event to your application. Verify its signature according to the provider’s documentation and associate the payment with the correct order. Then check the relevant payment status before releasing fulfillment. Receiving a message does not, by itself, prove that a payment succeeded.
Stripe also warns that events can arrive more than once and out of order. Make processing idempotent: receiving the same message again must not deduct stock again or trigger another shipment. Retain processed event IDs and protect the business operation itself, such as releasing a particular order only once. Checking event IDs alone does not cover every form of duplicate processing. The webhook documentation sets out these conditions.
At larger volumes, a queue can handle follow-up work. Acknowledge receipt only after the message has been durably recorded, then monitor whether processing succeeds. Otherwise, an invisible failure merely moves from checkout into the background.
Test exceptions and measure where orders stall
A successful test order establishes that one path works. For this campaign, I would also check at least the following situations in an appropriate test environment:
- A basket created before the promotion, checked out during and after it.
- Two concurrent requests for the last item.
- A failed payment followed by another attempt.
- A payment confirmation arriving late or twice.
- A temporary failure in the stock or order integration.
- The entire checkout on mobile and without a mouse.
For each situation, record what the customer should see and which status the order should reach. Error messages should help people recover without unnecessarily clearing their input. Dogteam’s article on accessible form validation explores that interaction further. These checks are a proposed test plan; no online store was load-tested and no checkout implementation was tested for this article.
During the promotion, monitor both sales and processing: paid orders, technical checkout errors, unprocessed payment events and growing delays. Reconcile payments against orders to find discrepancies. Use order and payment records to check revenue; browser analytics provide an additional view of behaviour in the funnel.
Decide beforehand who intervenes and when the promotion should be paused. Set thresholds that fit your normal traffic and recovery capacity. Buying more traffic while confirmed payments are waiting to be processed compounds the problem.
This is where technology contributes most to Black Friday: making the commercial promise deliverable. Start with one offer and follow it all the way to a correctly processed order. If that journey remains predictable when things go wrong, you have a much stronger foundation for scaling the campaign.
Sources checked on 10 November 2026. The technical examples use WooCommerce and Stripe documentation; other platforms have their own settings and guarantees.
Contact us
Tell us how we can help; we will respond within two working days.

